The objective of this proposed research is to develop a diagnosis-based classification system for long-term care patients which will be useful in predicting a patient's general potential for change in functional capabilities. The classification scheme will be developed using cluster analysis to group nursing home residents with similar patterns of diagnoses. The utility of the system will be evaluated by assessing its power to predict length of stay, discharge, mortality and changes in functional and mental status at a follow-up assessment. Three data sets which were previously collected as a part of research activities supported by the National Center for Health Services Research will be used to develop the classification scheme, test its stability across different samples of patients and evaluate its usefulness.